MTPY presents A.V.A
Atharva Virtual Automation
A modular, mechatronic humanoid platform that combines on-board intelligence, real-time control and precision actuation for human-grade interaction.
A.V.A is a distributed-architecture humanoid: Raspberry Pi runs perception, dialogue and web UI, Arduinos run deterministic safety and servo logic, and PCA9685 modules supply jitter-free PWM for synchronized multi-servo gestures.
Why Atharva Chose AVA
Who Benefits?
Students, educators, researchers and event demo teams benefit — AVA is built as an educational & demo platform that's easy to extend and service.
What is AVA?
A modular humanoid platform combining perception (camera/audio), AI decisioning, and precision actuation for lifelike gestures and interactions.
Why AVA Over Simpler Robots?
Because AVA separates brain, reflexes and actuation (Pi → Arduino → PCA9685), preventing single-point failures, eliminating servo jitter, and enabling context-aware gestures rather than pre-recorded loops.
When Should You Use AVA?
For public demos, classroom labs, research prototypes, and interactive kiosks where uptime, safety and expandability matter. It's designed to run long demos with reliable posture holding and safe fail-states.
Where Does AVA Fit?
As the bridge between hobby robots and research prototypes — ideal as a testbed before moving to industrial hardware or for on-site interactive deployments.
Which Components Matter Most?
The Pi (high-level AI & web UI), Arduinos (real-time safety & control), PCA9685 (hardware PWM) and dedicated buck converters (power stability). Each plays a defined role.
What Outcomes Will Atharva See?
Faster iterations (modular software stack), safer demos (separate power & watchdogs), and human-like gestures (synchronized PWM + motion planning) — measurable as fewer resets, smoother motion and easier maintenance.
Explore AVA
Software
Brain & UI
Runs perception, speech, ROS2 nodes, motion planning and the web dashboard. Uses TensorFlow-Lite/ONNX for lightweight models and supports Coral/AI HAT accelerators for heavier inference.
- →Pi runs ROS2 & dialogue
- →Middleware for PCA/Arduino orchestration
- →FastAPI/WebSocket web UI
Hardware
Motors, Power & Boards
Raspberry Pi (companion computer), Arduino Mega/Uno (real-time controllers), PCA9685 boards for PWM, and dedicated buck converters for stable servo rails — a pro mechatronics stack.
- →Pi5 recommended for performance
- →Separate upper/lower power rails
- →Thermal best practices
Integrated
System Level
A.V.A is built as a coordinated system (perception → decision → motion) rather than isolated gestures — enabling context-aware responses, safety interlocks and modular expansion.
- →Distributed brain architecture
- →Motion intelligence (gesture-level)
- →Modular expansion & OTA
Smooth Gestures
PCA9685 hardware PWM removes jitter.
Real-time Safety
Arduino watch-dogs keep posture if the Pi hangs.
Scalable
Add PCA boards and servos; designed to grow.
Dismantling & Service Walkthrough
Video player placeholder
A step-by-step dismantling video that demonstrates safe disassembly, connector mapping, how to replace servos, where to check fuses and how to reassemble for demo readiness.
00:00 — Intro & safety
Power disconnect procedure
01:10 — Head module removal
Camera & MG996R wiring
03:45 — Upper body PCA9685
Channel mapping to joints
06:20 — Lower body power
Buck converter & fuse checks